Proposal (#212) to South American Classification Committee

Split Gypopsitta from Pionopsitta

Effect on SACC: This proposal would remove all but one species from current Pionopsitta and place these in a separate genus, Gypopsitta.

Background: Traditional classifications (e.g., Peters 1937, Pinto 1937, Phelps & Phelps 1958a, Meyer de Schauensee 1970, Forshaw 1973) treated Pionopsitta vulturina in a separate monotypic genus, Gypopsitta, but Haffer (1974), Cracraft and Prum (1988), and others noted that vulturina was clearly a member of Pionopsitta on the basis of biogeography and plumage pattern. SACC classification treats vulturina as a member of Pionopsitta on this basis.

New information: Ribas et al. (2005) found, using mtDNA sequence data (2181 bp), that Pionopsitta as currently constituted is not monophyletic. The type species for the genus, pileata, is only distantly related to the others, which in fact are more closely related to, for example, Hapalopsittaca, Amazona, Graydidasculus, Pionus, and even Triclaria than to Pionopsitta pileata. Bootstrap and Bayesian posterior probabilities for the nodes in their tree are generally high. Rather than expand Pionopsitta to include a broad array of parrot genera, Ribas et al. (2005) resurrected Gypopsitta for the other species.

Analysis: Haffer (1974) and others long ago recognized that P. pileata was the "oddball" in the genus, and so the genetic data are not surprising in corroborating this. What is surprising is how distant P. pileata is from the Gypopsitta group, and how far one would have to expand Pionopsitta to retain it as a monophyletic group, e.g., minimally all the genera mentioned above. Although the genetic analysis covered only two mtDNA genes, it is difficult to imagine that mtDNA would be this misleading at this level (intergeneric) of analysis. Comforting is that formal analysis of plumage characters (Cracraft and Prum 1988) shows that pileata is the outgroup to the rest; in fact, with the benefit of hindsight, if you look at specimens, you wonder why pileata was ever associated with the Gypopsitta group. [Perhaps someone could check Bret et al.'s parrot CD for vocal information.]

Keeping Pionopsitta and Gypopsitta adjacent in a linear sequence misrepresents their degree of relatedness, but major fiddling with the linear sequence at this point seems inadvisable until we have better data on the relationships among these parrot genera. However, Ribas et al.'s genetic data strongly support Hapalopsittaca amazonina as the sister taxon to Gypopsitta (H. melanotis not sampled). The plumage similarities between Hapalopsittaca and Gypopsitta/Pionopsitta are reflected in historical classifications of their component species (see SACC Notes) and Cracraft and Prum's use of H. melanotis as the outgroup taxon for their plumage analysis of Gypopsitta/Pionopsitta. Therefore, I recommend that our sequence go Gypopsitta-Hapalopsittaca-Pionopsitta to emphasize that there is no evidence that Gypopsitta and Pionopsitta are sisters, but with minimum disturbance to the linear sequence until more data are available.

Recommendation: YES. There are no data of which I am aware that support the continued association of Pionopsitta pileata with the Gypopsitta group, and so to continue to treat them as congeners would be highly misleading.

Literature Cited (see SACC Biblio for the rest):

CRACRAFT, J., AND R. O. PRUM. 1988. Patterns and processes of diversification: speciation and historical congruence in some Neotropical birds. Evolution 42: 603-620.

RIBAS, C. C., R. GABAN-LIMA, C. Y. MIYAKI, AND J. CRACRAFT. 2005. Historical biogeography and diversification within the Neotropical parrot genus Pionopsitta (Aves: Psittacidae). Journal Biogeography 32:1409-1427.


V. Remsen, March 2006

===========================================================================

Comments from Stiles: "YES. The genetic (as well as morphological) data clearly mandate this change, supported by two different studies and several genes (pity that Gypopsitta is a bit misleading as a name for most species, but priority rules!)."

Comments from Robbins: "YES. The Ribas et al. (2205) genetic data and their taxonomic recommendations of resurrecting Gypopsitta seem solid. I vote 'yes'."

Comments from Nores: "YES. Es muy claro en el análisis de Ribas et al. (2005) que Pionopsitta pileata está muy poco emparentada con el resto de las especies que proponen poner en el género Gypopsitta. Esto también resulta evidente cuando se comparan las especies desde el punto de vista morfólógico, principalmente la coloración."

Comments from Zimmer: "YES. The evidence presented is convincing. As Gary notes, it is unfortunate that priority dictates the name Gypopsitta, which is great for two of the species but doesn't hold for the others. In response to Van's query about vocalizations of pileata versus the other species currently in Pionopsitta: pileata is quite different from all of the others with which I'm familiar, whereas vulturina, aurantiocephala, barrabandi and caica all have some similarities with one another."

Comments from Jaramillo: "YES - Change clear based on genetic data and morphological data, also unpublished vocal data fide Zimmer."

Comments from Pacheco: "YES. Os dados disponibilizados por Ribas  et al. (2005) são plenamente convincentes para a adoção do arranjo proposto.

"É oportuno mencionar que um trabalho independente chegou a conclusões semelhantes:

"Eberhard, J. E. & Bermingham, E. 2005. Phylogeny and comparative biogeography of Pionopsitta parrots and Pteroglossus toucans. Mol. Phyl. & Evol. 36: 288304."